From patchwork Wed May 22 22:32:34 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Stubbs X-Patchwork-Id: 1103712 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=gcc.gnu.org (client-ip=209.132.180.131; helo=sourceware.org; envelope-from=gcc-patches-return-501486-incoming=patchwork.ozlabs.org@gcc.gnu.org; receiver=) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=codesourcery.com Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; unprotected) header.d=gcc.gnu.org header.i=@gcc.gnu.org header.b="Ze9146rD"; dkim-atps=neutral Received: from sourceware.org (server1.sourceware.org [209.132.180.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 458S7k4Hpwz9s1c for ; Thu, 23 May 2019 08:32:50 +1000 (AEST) DomainKey-Signature: a=rsa-sha1; c=nofws; d=gcc.gnu.org; h=list-id :list-unsubscribe:list-archive:list-post:list-help:sender:from :subject:to:message-id:date:mime-version:content-type; q=dns; s= default; b=dJyfIL/PGOg47+xgYloANRhLDChkEGR7EBuyXgQINMMrx1B5eZBn5 aTT/I8gqybEo6T/jjEgWjsezvgS4yMfcJ/Jw4Uy80I1puQFC9nViWXkyoJ6xfop8 wnXfVp3lWJ5tCNeudzzuk4/UFdT8AhNz7x2AeM3XeCGgtmId13aeoU= DKIM-Signature: v=1; a=rsa-sha1; c=relaxed; d=gcc.gnu.org; h=list-id :list-unsubscribe:list-archive:list-post:list-help:sender:from :subject:to:message-id:date:mime-version:content-type; s= default; bh=d1DB2FJrU7AWe3RULBjie9WfnhA=; b=Ze9146rDmrVrl7qhRKpm 1Q52LDcfQt+oJ24wx5FDGQSfeOkeb3HsW8KNvNdMvvHgQR1SN+WdeBQAHSNZT0G6 AA1LS8PeX/hBqdhBLSYfN6mOMSOqb/Q687sqYA/mUqh1lfOf0WCmVy++M4Myct8r EgPZDqmHkrlSUFLNiDg6KyA= Received: (qmail 50460 invoked by alias); 22 May 2019 22:32:43 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Delivered-To: mailing list gcc-patches@gcc.gnu.org Received: (qmail 50445 invoked by uid 89); 22 May 2019 22:32:43 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-16.7 required=5.0 tests=AWL, BAYES_00, GIT_PATCH_0, GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, KAM_SHORT, KAM_STOCKGEN, RCVD_IN_DNSWL_NONE, SPF_PASS autolearn=ham version=3.3.1 spammy=monitor X-HELO: relay1.mentorg.com Received: from relay1.mentorg.com (HELO relay1.mentorg.com) (192.94.38.131) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Wed, 22 May 2019 22:32:41 +0000 Received: from nat-ies.mentorg.com ([192.94.31.2] helo=svr-ies-mbx-01.mgc.mentorg.com) by relay1.mentorg.com with esmtps (TLSv1.2:ECDHE-RSA-AES256-SHA384:256) id 1hTZmp-0001RA-Ao from Andrew_Stubbs@mentor.com for gcc-patches@gcc.gnu.org; Wed, 22 May 2019 15:32:39 -0700 Received: from [127.0.0.1] (137.202.0.90) by svr-ies-mbx-01.mgc.mentorg.com (139.181.222.1) with Microsoft SMTP Server (TLS) id 15.0.1320.4; Wed, 22 May 2019 23:32:35 +0100 From: Andrew Stubbs Subject: [committed, amdgcn] Add support for constructors and destructors To: "gcc-patches@gcc.gnu.org" Message-ID: Date: Wed, 22 May 2019 23:32:34 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.6.1 MIME-Version: 1.0 This patch adds support for constructors that run before "main", and destructors that run afterwards. This is prerequisite for supporting more of libgfortran (and other libraries, including libstdc++). The init_array and fini_array sections are created, and gcn-run is modified to launch the initialization and finalization kernels, in addition to the main entry point. Add support for constructors and destuctors on GCN 2019-05-22 Kwok Cheung Yeung Andrew Stubbs AMDGPU-230 gcc/ * config.gcc (gcc_cv_initfini_array): Set for AMD GCN. * config/gcn/gcn-run.c (init_array_kernel, fini_array_kernel): New. (kernel): Rename to... (main_kernel): ... this. (load_image): Load _init_array and _fini_array kernels. (run): Add argument for kernel to run. (main): Run init_array_kernel before main_kernel, and fini_array_kernel after. * config/gcn/gcn.c (gcn_handle_amdgpu_hsa_kernel_attribute): Allow amdgpu_hsa_kernel attribute on functions. (gcn_disable_constructors): Delete. (TARGET_ASM_CONSTRUCTOR, TARGET_ASM_DESTRUCTOR): Delete. * config/gcn/crt0.c (size_t): Define. (_init_array, _fini_array): New. (__preinit_array_start, __preinit_array_end, __init_array_start, __init_array_end, __fini_array_start, __fini_array_end): Declare weak references. diff --git a/gcc/config.gcc b/gcc/config.gcc index 76bb316942d..7122c8ed1c8 100644 --- a/gcc/config.gcc +++ b/gcc/config.gcc @@ -1424,6 +1424,8 @@ amdgcn-*-amdhsa) extra_programs="${extra_programs} mkoffload\$(exeext)" tm_file="${tm_file} gcn/offload.h" fi + # Force .init_array support. + gcc_cv_initfini_array=yes ;; moxie-*-elf) gas=yes diff --git a/gcc/config/gcn/gcn-run.c b/gcc/config/gcn/gcn-run.c index 00a71014c20..84718f42846 100644 --- a/gcc/config/gcn/gcn-run.c +++ b/gcc/config/gcn/gcn-run.c @@ -66,7 +66,9 @@ bool debug = false; hsa_agent_t device = { 0 }; hsa_queue_t *queue = NULL; -uint64_t kernel = 0; +uint64_t init_array_kernel = 0; +uint64_t fini_array_kernel = 0; +uint64_t main_kernel = 0; hsa_executable_t executable = { 0 }; hsa_region_t kernargs_region = { 0 }; @@ -427,14 +429,30 @@ load_image (const char *filename) XHSA (hsa_fns.hsa_executable_freeze_fn (executable, ""), "Freeze GCN executable"); - /* Locate the "main" function, and read the kernel's properties. */ + /* Locate the "_init_array" function, and read the kernel's properties. */ hsa_executable_symbol_t symbol; + XHSA (hsa_fns.hsa_executable_get_symbol_fn (executable, NULL, "_init_array", + device, 0, &symbol), + "Find '_init_array' function"); + XHSA (hsa_fns.hsa_executable_symbol_get_info_fn + (symbol, HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_OBJECT, &init_array_kernel), + "Extract '_init_array' kernel object kernel object"); + + /* Locate the "_fini_array" function, and read the kernel's properties. */ + XHSA (hsa_fns.hsa_executable_get_symbol_fn (executable, NULL, "_fini_array", + device, 0, &symbol), + "Find '_fini_array' function"); + XHSA (hsa_fns.hsa_executable_symbol_get_info_fn + (symbol, HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_OBJECT, &fini_array_kernel), + "Extract '_fini_array' kernel object kernel object"); + + /* Locate the "main" function, and read the kernel's properties. */ XHSA (hsa_fns.hsa_executable_get_symbol_fn (executable, NULL, "main", device, 0, &symbol), "Find 'main' function"); XHSA (hsa_fns.hsa_executable_symbol_get_info_fn - (symbol, HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_OBJECT, &kernel), - "Extract kernel object"); + (symbol, HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_OBJECT, &main_kernel), + "Extract 'main' kernel object"); XHSA (hsa_fns.hsa_executable_symbol_get_info_fn (symbol, HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_KERNARG_SEGMENT_SIZE, &kernarg_segment_size), @@ -684,7 +702,7 @@ gomp_print_output (struct kernargs *kernargs, bool final) /* Execute an already-loaded kernel on the device. */ static void -run (void *kernargs) +run (uint64_t kernel, void *kernargs) { /* A "signal" is used to launch and monitor the kernel. */ hsa_signal_t signal; @@ -822,11 +840,17 @@ main (int argc, char *argv[]) kernargs->heap_ptr = (int64_t) &kernargs->heap; kernargs->heap.size = heap_size; + /* Run constructors on the GPU. */ + run (init_array_kernel, kernargs); + /* Run the kernel on the GPU. */ - run (kernargs); + run (main_kernel, kernargs); unsigned int return_value = (unsigned int) kernargs->output_data.return_value; + /* Run destructors on the GPU. */ + run (fini_array_kernel, kernargs); + unsigned int upper = (return_value & ~0xffff) >> 16; if (upper == 0xcafe) printf ("Kernel exit value was never set\n"); diff --git a/gcc/config/gcn/gcn.c b/gcc/config/gcn/gcn.c index 6820837c763..47630c6edb4 100644 --- a/gcc/config/gcn/gcn.c +++ b/gcc/config/gcn/gcn.c @@ -305,9 +305,7 @@ static tree gcn_handle_amdgpu_hsa_kernel_attribute (tree *node, tree name, tree args, int, bool *no_add_attrs) { - if (FUNC_OR_METHOD_TYPE_P (*node) - && TREE_CODE (*node) != FIELD_DECL - && TREE_CODE (*node) != TYPE_DECL) + if (!FUNC_OR_METHOD_TYPE_P (*node)) { warning (OPT_Wattributes, "%qE attribute only applies to functions", name); @@ -3165,20 +3163,6 @@ gcn_valid_cvt_p (machine_mode from, machine_mode to, enum gcn_cvt_t op) || (to == DFmode && (from == SImode || from == SFmode))); } -/* Implement both TARGET_ASM_CONSTRUCTOR and TARGET_ASM_DESTRUCTOR. - - The current loader does not support running code outside "main". This - hook implementation can be replaced or removed when that changes. */ - -void -gcn_disable_constructors (rtx symbol, int priority __attribute__ ((unused))) -{ - tree d = SYMBOL_REF_DECL (symbol); - location_t l = d ? DECL_SOURCE_LOCATION (d) : UNKNOWN_LOCATION; - - sorry_at (l, "GCN does not support static constructors or destructors"); -} - /* }}} */ /* {{{ Costs. */ @@ -5995,10 +5979,6 @@ print_operand (FILE *file, rtx x, int code) #define TARGET_ARG_PARTIAL_BYTES gcn_arg_partial_bytes #undef TARGET_ASM_ALIGNED_DI_OP #define TARGET_ASM_ALIGNED_DI_OP "\t.8byte\t" -#undef TARGET_ASM_CONSTRUCTOR -#define TARGET_ASM_CONSTRUCTOR gcn_disable_constructors -#undef TARGET_ASM_DESTRUCTOR -#define TARGET_ASM_DESTRUCTOR gcn_disable_constructors #undef TARGET_ASM_FILE_START #define TARGET_ASM_FILE_START output_file_start #undef TARGET_ASM_FUNCTION_PROLOGUE diff --git a/libgcc/config/gcn/crt0.c b/libgcc/config/gcn/crt0.c index c0b93534eed..9b58029816c 100644 --- a/libgcc/config/gcn/crt0.c +++ b/libgcc/config/gcn/crt0.c @@ -19,5 +19,61 @@ see the files COPYING3 and COPYING.RUNTIME respectively. If not, see . */ +typedef long long size_t; + /* Provide an entry point symbol to silence a linker warning. */ void _start() {} + +#ifdef USE_NEWLIB_INITFINI + +extern void __libc_init_array (void) __attribute__((weak)); +extern void __libc_fini_array (void) __attribute__((weak)); + +__attribute__((amdgpu_hsa_kernel ())) +void _init_array() +{ + __libc_init_array (); +} + +__attribute__((amdgpu_hsa_kernel ())) +void _fini_array() +{ + __libc_fini_array (); +} + +#endif + +/* These magic symbols are provided by the linker. */ +extern void (*__preinit_array_start []) (void) __attribute__((weak)); +extern void (*__preinit_array_end []) (void) __attribute__((weak)); +extern void (*__init_array_start []) (void) __attribute__((weak)); +extern void (*__init_array_end []) (void) __attribute__((weak)); +extern void (*__fini_array_start []) (void) __attribute__((weak)); +extern void (*__fini_array_end []) (void) __attribute__((weak)); + +__attribute__((amdgpu_hsa_kernel ())) +void _init_array() +{ + /* Iterate over all the init routines. */ + size_t count; + size_t i; + + count = __preinit_array_end - __preinit_array_start; + for (i = 0; i < count; i++) + __preinit_array_start[i] (); + + count = __init_array_end - __init_array_start; + for (i = 0; i < count; i++) + __init_array_start[i] (); +} + +__attribute__((amdgpu_hsa_kernel ())) +void _fini_array() +{ + size_t count; + size_t i; + + count = __fini_array_end - __fini_array_start; + for (i = count; i > 0; i--) + __fini_array_start[i-1] (); +}